INDEX

Explanations

4

This neuron fires on tokens that occur as section or article titles (e.g. headings and sidebar/related-posts titles).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 மற்றும்

1.00

 ০৯

0.75

)/

0.74

 וא

0.74

 ডাঃ

0.72

(${

0.71

 για

0.68

 denoted

0.68

 ולא

0.68

 পশ্চিম

0.67

POSITIVE LOGITS

這些

0.75

immediate

0.73

йдёт

0.73

这些

0.72

йдет

0.72

adequate

0.72

ల్ని

0.71

joyed

0.70

 అలా

0.69

je

0.68

Activations Density 0.001%